Coded Conversations: A Critical Discourse Analysis of Algospeak as a Subversive Linguistic Practice on X

Abstract

Platforms decide which words are too risky to publish, and users learn to speak around them. This study examines algospeak, the deliberate misspelling, symbol substitution, and re-coding of words to evade automated moderation, as a subversive linguistic practice on X (formerly Twitter). Using Critical Discourse Analysis (Wodak & Meyer, 2022) and platform studies scholarship (Gorwa et al., 2020), it analyzes 25 English-language posts from January 2025 to May 2026, sampled purposively across mental health discourse, political speech, meta-discourse on moderation, and emoji-based encoding. Six strategies emerged: lexical substitution, phonetic distortion, graphic and numeric distortion, acronymic encoding, subcultural lexicons, and emoji displacement. The findings show that the burden of encoding falls unevenly on users discussing mental health, political marginality, and geopolitically sensitive topics, and that many users treat algospeak as a conscious act of resistance. The study recommends context-sensitive moderation, greater transparency, and algospeak literacy in digital education.